BACKGROUND: Agriculture employs 10% of Wisconsin workers and has the third highest reported annual rate of workplace injury in the state. Agricultural injuries may be underreported due to limitations of current surveillance systems. Our Minnesota Collaborative partner used alternative data sources to improve agricultural injury surveillance (Landsteiner AMK et al., 2015). We applied this approach to Wisconsin data to pilot a potentially new agricultural injury indicator.
METHODS: We described agricultural injuries in Wisconsin hospital discharge and emergency department (ED) datasets during 2005-2014 using seven external cause of injury codes (ICD-9-CM, E codes) to identify distinct groups of probable and possible agricultural injuries. Rate denominators of individuals living on a farm plus individuals working but not living on a farm were calculated using the American Community Survey (ACS; US Census Bureau) and the Census of Agriculture (US Department of Agriculture).
RESULTS: Of 28,381 agricultural injuries identified, 11% were from inpatient records and 89% were ED records. Thirteen percent of agricultural injuries were considered probable, while 87% met less stringent criteria and were classified as possible. The most frequently coded injuries were from riding animals (E828) and other animal-caused injuries (E906.8). Annual rates for probable cases ranged from 1.5–6.9/1000 workers; possible case rates ranged from 14.8–18.8/1000. Although workers aged 35–64 years incurred the most injuries, workers aged ≥65 years were more likely to be probable cases. Probable cases were predominantly male (85%) and possible cases were predominantly female (62%). Two-thirds of injuries occurred during May through October, and 74% were incurred by residents of rural counties. Potential for underreporting was found in several parts of this analysis. Workers’ Compensation was an anticipated primary or secondary payer in only 6% of all reported injuries. Use of Wisconsin hospital and ED data resulted in nearly a threefold increase in the number of identified injuries compared to the Survey of Occupational Injury and Illness (SOII) (e.g., 10-year average of 2838 injuries/year vs. SOII’s 2015 estimate of 1000 injuries/year).
CONCLUSIONS: Our results were largely consistent with Minnesota’s findings. Despite the limitations of hospital and ED discharge data, the proposed indicator provides a more sensitive measure of agricultural injury in Wisconsin and has potential for use in other agricultural states.